Decision making And GeNerative AI: benefits and limitations
What is generative AI, how can it be used for decision making, to accelerate learning, and what are some current limitations?
Generative Artificial Intelligence
Generative AI is a type of artificial intelligence (AI) that uses algorithms to create content, such as images, music, or text. It does this by applying rules coded in the algorithms to an existing set of data. For example, the language model ChatGPT was trained using text from sources such as Wikipedia, news articles, books, and scientific papers. Given these sources, we can prompt ChatGPT and the program then generates a response based on patterns of consensus that exist within the dataset.
Better Decisions
It is this ability of Generative AI to use the dataset to quickly gather relevant information and develop multiple options, that is a major benefit to decision makers. With language models, like ChatGPT, decision makers can ask additional or follow up questions in order to refine the results. Decision makers can use the AI to seek feedback and explore consequences of the various options.
Accelerated Learning
Another benefit of Generative AI is that it can be used to help accelerate learning. This is possible by using AI to quickly generate new and diverse training scenarios. Students can gain experience faster by navigating a range of simulated environments where they can develop their skills in a safe and controlled setting. And because Generative AI responds to prompts, it can help to create better engagement with students as they receive individualized feedback. This means, that both novices and experts can interact with a simulation and receive feedback that corresponds to their skill level.
Limitations
For all of the benefits of Generative AI, there are also some drawbacks to consider.
First, one of the main limitations is that AI is only as good as the available data. The popular adage, garbage in, garbage out, applies to Generative AI. When the consensus data is wrong or outdated, then most likely the generated content will reflect these same errors.
Second, if used as a substitute for human judgment and expertise, Generative AI can lead to worse decisions. While the generated content can provide a foundation for discovery, it doesn’t replace the nuanced decision-making abilities of an expert.
Third, over reliance on Generative AI may lead to worse learning outcomes. It could potentially reduce opportunities for students to grasp the underlying fundamentals of a problem or to develop the skills necessary to navigate real world situations. And a long term consequence, is that over time the pool of expertise available to create a reliable data set might be reduced, which in turn will reduce the quality of the content generated.
Conclusions
In summary, Generative AI is a useful tool that can be used many different ways, including to help inform decision making and to accelerate learning. At the same time, it is not without certain limitations. Therefore, like any other tool, we need to learn how to use it appropriately. We need to use the right tool for the right job. In this sense, there is a balance that we need to find, using Generative AI as a tool to try and enhance, rather than replace human expertise.